Within the the Citoid logstash dashboard there is data showing the top errors by requested domains.
This task seeks to help us better understand URL requests made through Citoid by calculating the overall rate at which requests of this sort fail.
Longer term, we'd like to be able to calculate the failure/success rates by other requests (e.g. ISBN, DOI, PubMed ID, etc.). Tho, doing so would require instrumentation we've not yet added. Ticket needed
Knowing the above will help the Editing Team ensure we are allocating our attention to addressing the most prevalent failures.
Requirements
- Data showing rates at which Citoid URL requests fail and succeed within the following time intervals:
Last 24 hours
Excluding PDFs:
All formats: 80.6% success
No zotero format: 86.8% success
Only zotero format: 79.0 % success
Including PDFs:
794 urls ending in .pdf out of 93785+794 requests total= .8% of requests.
https://logstash.wikimedia.org/goto/5c8ad02c549621b486fbe96095109a82
Last 1 week
Excluding PDFs
All formats: 76.4 % success
No zotero format: 86.2% success
Only zotero format: 75% success
Last month
Excluding PDFs
All formats: 76.1%
No zotero format: 87.3%
Only zotero format: 74.7%
All formats an input type, including pdfs (metrics):
83.7 % success
Last 6 months
TODO
- Data "1." describes needs to be filter-able by request origin (Wikimedia and External/3rd Party)